The majority of currently-approved small molecule drugs are absorbed across the mucosa of the small intestine to provide delivery to the systemic circulation. In fact, small molecule drugs are selected based upon their stability and efficient absorption across intestinal mucosae. A similar oral delivery of biologically-active polypeptides (referring to a polymer composed of amino acid residues; typically defined as a protein or peptide) has been a long-standing goal of the pharmaceutical industry. As the gastrointestinal (GI) tract is designed to digest dietary proteins and peptides, there are numerous physical, physiological, and biological barriers that limit the feasibility of therapeutic proteins and peptides uptake from the intestine in a manner similar to that achievable with small molecules; Mahato, R. I., et al., Crit Rev Ther Drug Carrier Syst, 20(2-3):p. 153-214 (2003).
A number of technologies have been identified that can be used to protect therapeutic proteins and peptides through the stomach, allowing them to reach the absorptive surface of epithelial cells in the small intestine and separating them from the gastric and intestinal environments that function to destroy dietary proteins and peptides. Unfortunately, however, the efficient transport across this simple, single layer of cells remains a substantial barrier due to the intracellular trafficking to destructive lysosome compartments after endosomal uptake of polypeptides at the luminal surface; Woodley, J. F., Crit Rev Ther Drug Carrier Syst, 11(2-3):p. 61-95 (1994). Indeed, this barrier is designed to inhibit uptake of proteins and peptides until these macromolecules can be sufficiently degraded for absorption through amino acid and di- or tri-peptide transporters. In this regard, a number of efforts have been examined to overcome the physical, physiological, and biological barriers of the intestinal mucosae.
There are two basic routes across the simple epithelium that constitutes the cellular barrier of the intestinal mucosae. Specifically, once across the covering mucus layer, a molecule could move between adjacent epithelial cells (paracellular route) or move through cells (transcellular route) via a series of vesicles that traffic within, but do not mingle, contents with the cytoplasm; T. Jung et al., Eur J Pharm Biopharm, 50:147-160 (2000). In other words, in both routes, a transport protein or peptide therapeutic does not enter into the cell but rather stays in an environment external to the cell's cytoplasm.
The primary barrier to casual movement of therapeutic protein and peptide movement through the paracellular route is a complex of proteins at the apical neck of these cells known as the tight junction (TJ). While transient opening and closing of TJ structures can facilitate transport of peptides across intestinal epithelia, this approach has key limitations: e.g., it does not work well for molecules above ˜5 kDa; it has the potential for non-selective entry of materials into the body from the intestinal lumen; and it represents a route that involves only a small fraction of the surface area of the intestinal epithelium.
The primary barrier to casual migration of protein or peptide therapeutics across cells via the transcellular route is a default vesicle trafficking that delivers the contents of these vesicles to a destructive (lysosomal) pathway. As compared to the paracellular route, movement through the vesicular transcellular route can accommodate materials as large as 100 nm in diameter, involves essentially the entire epithelial cell surface, and can be highly selective in uptake of materials through the use of receptor-ligand interactions for vesicle entry. Thus, the transcellular route is very appealing for the epithelial transport of protein or peptide therapeutics if the destructive pathway can be avoided.
Some pathogens have solved the trafficking barrier problem, as demonstrated by the efficient transcytosis of secreted polypeptide virulence factors which function to facilitate and/or stabilize infection of a host. Exotoxins represent a class of proteins released by a variety of microorganisms which function as potent virulence factors. Exotoxins function on multi-cellular organisms with the capacity to acts as potent toxins in man; Roszak, D. B., and Colwell, R. R., Microbiol Rev 51:365-379 (1987). These proteins commonly kill or inactivate host cells through mechanisms that involve selective disruption of protein synthesis. Accordingly, only a few molecules are required to kill, consistent with the observation that bacterial exotoxins are some of the most toxic agents known. A subset of these proteins comprised of the family of proteins that consists of diphtheria toxin (DT) from Corynebacterium diphtheria, exotoxin A from Pseudomonas aeruginosa (PE), and a recently identified protein termed Cholix from Vibrio cholera function to intoxicate host cells via the ADP-ribosylation of elongation factor 2 (EF2); Yates, S. P., et al., Trends Biochem Sci, 31:123-133 (2006). These exotoxins are synthesized as a single chain of amino acids that fold into distinct domains that have been identified as having specific functions in targeting, entry, and intoxication of host cells.
The biology of exotoxin A from Pseudomonas aeruginosa (PE) has recently been described; Mrsny, R. J., et al., Drug Discov Today, 7(4): p. 247-58 (2002). PE is composed of a single chain of 613 amino acids having a theoretical molecular weight (MW) of 66828.11 Da, an isoelectric point (pI) of 5.28, and that functionally folds into three discrete domains, denoted domain I (Ala1-Glu252), domain II (Gly253-Asn364), domain III (Gly405-Lys613, and which contains a ADP-ribosyltransferase activity site), and a short disulfide-linked loop linking domains II and III which is known as the Ib loop (Ala365-Gly404). The organization of these domains at pH 8.0 have determined from crystal diffraction at a resolution of ˜1.5 Å; Wedekind, J. E. et al., J Mol Biol, 314:823-837 (2001). Domain I (Ia+Ib) has a core formed from a 13-stranded β-roll, domain II is composed of six α-helices, and domain III has a complex α/β-folded structure. Studies have supported the idea that the modular nature of PE allows for distinct domain functions: domain I binds to host cell receptors, domain II is involved in membrane translocation, and domain III functions as an ADP-ribosyltransferase. It appears that PE is secreted by P. aeruginosa in close proximity to the epithelial cell apical surface, possibly in response to environmental cues and/or cellular signals; Deng, Q. and J. T. Barbieri, Annu Rev Microbiol, 62:p. 271-88 (2008). Once secreted, internalization into cells occurs after domain I of PE binds to the membrane protein α2-macroglobulin, a protein which is also known as the low-density lipoprotein receptor-related protein 1 (LRP1) or CD91; see, e.g., FitzGerald, D. J., et al., J Cell Biol, 129(6):p. 1533-41 (1995); Kounnas, M. Z., et al., J Biol Chem, 267(18): p. 12420-3 (1992). Following internalization, PE avoids trafficking to the lysosome and is instead efficiently delivered to the basolateral surface of the cell where it is released in a biologically-active form; Mrsny, R. J., et al., Drug Discov Today, 7(4): p. 247-58 (2002). Once across the epithelium, PE functions as a virulence factor by entering into CD91-positive cells within the submucosal space (macrophage and dendritic cells) where it then intersects with an unfolding pathway that leads to the cytoplasmic delivery of domain III; see, e.g., Mattoo, S., Y. M. Lee, and J. E. Dixon, Curr Opin Immunol, 19(4): p. 392-401 (2007); Spooner, R. A., et al., Virol J, 3: p. 26 (2006).
Vibrio cholerae bacterium is best known for its eponymous virulence agent, cholera toxin (CT), which can cause acute, life-threatening massive watery diarrhea. CT is a protein complex composed of a single A subunit organized with a pentamer of B subunits that binds to cell surface GM1 ganglioside structures at the apical surface of epithelia. CT is secreted by V. cholera following horizontal gene transfer with virulent strains of V. cholerae carrying a variant of lysogenic bacteriophage called CTXf or CTXφ. Recent cholera outbreaks, however, have suggested that strains of some serogroups (non-O1, non-O139) do not express CT but rather use other virulence factors. Detailed analyses of non-O1, non-O139 environmental and clinical data suggested the presence of a novel putative secreted exotoxin with some similarity to PE.
Jorgensen, R. et al., J Biol Chem, 283(16):10671-10678 (2008) reported that some strains of V. cholerae did, in fact, contain a protein toxin having similarity to PE and which they termed Cholix toxin (Cholix). Compared to PE, Cholix has a slightly larger theoretical MW (70703.89 Da) and a slightly more acidic theoretical pI (5.12). The crystal structure of the 634 amino acid Cholix protein has been resolved to ˜2 Å. The domain structure and organization was found to be somewhat similar to PE: domain I (Val1-Lys265), domain II (Gly266-Ala386), domain III (Arg426-Lys634), and a Ib loop (Ala387-Asn425). Additional structural similarity to PE includes: a furin protease site for cellular activation; a C-terminal KDEL sequence that can route the toxin to the endoplasmic reticulum of the host cell; and an ADP-ribosyltransferase activity site within domain III.
Remarkably, PE and Cholix share no significant genetic and limited similarity by amino acid alignment. Searching the genome of V. cholera for PE-like nucleotide sequences fails to result in a match of any kind. It is only at the protein sequence level is there the hint that an PE-like protein could be produced by this bacterium. Even here, there is only a 32% homology between the amino acid sequences of PE and Cholix with similarities (42% homology) being focused in the ADP ribosylation elements of domain III, and with low levels of amino acid homology (˜15-25%) for most segments of domains I and II for the two proteins. Moreover, this overall arrangement of Cholix relative to PE is even more striking since these two proteins with similar elements were derived from two distinct directions: P. aeruginosa is a GC-rich bacterium while V. cholera is AT-rich. That these two toxins evolved from two different genetic directions to arrive at nearly the same structure but with only 32% amino acid homology suggests that structural and functional similarities of Cholix and PE are likely based upon similar survival pressures rather than through similar genetic backgrounds. The very low amino acid homology of domains I and II for these two proteins stress the functional importance of the folded structures of these two proteins and not their amino acid sequences.
The C-terminal portion of Cholix and PE appear to function in the intoxication of cells through ADP-ribosylation of EF2 in comparable ways. Recent studies where the latter half of Cholix (domain I deleted) targeted to cancer cells through conjugation to an antibody directed to the transferrin receptor suggests that the C-terminal portions of PE and Cholix involved in ADP-ribosylation of EF2 are indeed functionally similar; Sarnovsky, R., et al., Cancer Immunol Immunother 59:737-746 (2010). While this distal portion of Cholix is 36% identical and 50% similar to PE, polyclonal antisera raised in animals as well as sera from patients having neutralizing immune responses to this same distal portion of PE failed to cross-react with this latter portion of Cholix. Similarly, antisera raised to this Cholix failed to cross-react with PE. This data suggests that while both PE and Cholix share a capacity to intoxicate cells through a similar mechanism and that these two proteins share a common core structure, there are striking differences in their elements that are expressed at the surface of these proteins.
As previous studies using PE have demonstrated that this toxin readily transports across polarized monolayers of epithelial cells in vitro and in vivo without intoxication; Mrsny, R. J., et al., Drug Discov Today, 7(4): p. 247-58 (2002), the present inventors have commenced research to further evaluate the properties and biology of Cholix, with a particular focus on the functional aspects of the proximal portions of Cholix; specifically, the use of domains I and II to facilitate transport across intestinal epithelial monolayers. As domains I and IIa appeared to be the only essential elements of PE required for epithelial transcytosis, it was particularly important to examine these same domains in Cholix. As stated previously, there is only ˜15%-25% amino acids homology over most of the regions that would be considered to be part of domains I and IIa. The present inventors examined the domains though a series of studies: monitoring the biological distribution of Cholix following application to epithelial surfaces in vivo, assessment of Cholix transcellular transport characteristics across polarized epithelial cell monolayers in vitro, and delivery of a biologically-active cargo genetically integrated into the Cholix protein at its C-terminus. Preliminary data generated by genetically fusing the first two domains of Cholix (amino acids 1-386) to green fluorescent protein (GFP) or chemically coupling these expressed domains to 100 nm diameter latex beads demonstrated that Cholix attached to 100 nm latex beads were observed to transport across intestinal epithelial monolayers in vitro and in vivo. That the GFP cargo retained its fluorescent character during and after the transcytosis process also support the contention that Cholix utilizes a non-destructive (or privileged) trafficking pathway through polarized epithelial cells. This outcome bodes well for its (repeated) application as a tool to deliver biologically active cargos across epithelial barriers of the body, such as those in the respiratory and gastrointestinal tracts.
Also of important note from the preliminary studies is the observation which suggests an apparent cell receptor interaction difference between PE and Cholix. As stated previously, PE enters into epithelial cells after domain I of PE binds to the membrane protein α2-macroglobulin, a protein which is also known as the low-density lipoprotein receptor-related protein 1 (LRP1) or CD91. While the exact identity of the surface receptor for Cholix has not been established, preliminary studies suggest that Cholix does not intoxicate some cell lines that express CD91 but intoxicates some cell lines that do express CD91. It is currently unclear what other receptors, beyond CD91, might be involved epithelial transcytosis of PE. Nevertheless, Cholix and PE appear to have distinct cell receptor interactions, demonstrating clear differences that are sufficient to suggest very different and unanticipated applications for both oral biologics and the intracellular delivery of bioactive agents.